The Network Antidote: An Agent-Based Model of Discrimination in Labor Markets

• In Event: 272 - Regular Session. Rational Choice

Abstract

This study shows that large inequality in employment can be pervasive in the labor market even when there are no differences in average quality between different worker categories. By simulating hiring choices of impartial employers through an agent-based model, we show that social networks have a different impact on discrimination depending on which type of network ties we consider. While the frequency of worker referrals has a U-shaped effect on the extent of discrimination, business recommendations do not have a prominent impact. We find that certain social network mechanisms might decrease discrimination compared to a market composed of atomized employers as networks can allow employers to overcome individual sampling biases. We demonstrate that ties between workers and employers that are characterized by their affective content decrease discrimination more than referrals that are characterized purely by information exchange.

Authors

• Károly Takács, Corvinus University-Budapest

• Flaminio Squazzoni, University of Brescua

• Giangiacomo Bravo, Linnaeus University